A methodological review of the high-dimensional propensity score in comparative-effectiveness and safety-of-interventions research finds incomplete reporting relative to algorithm development and robustness

OBJECTIVES

The use of secondary databases has become popular for evaluating the effectiveness and safety of interventions in real-life settings. However, the absence of important confounders in these databases is challenging. To address this issue, the high-dimensional propensity score (hdPS) algorithm was developed in 2009. This algorithm uses proxy variables for mitigating confounding by combining information available across several healthcare dimensions. This study assessed the methodology and reporting of the hdPS in comparative effectiveness and safety research.

STUDY DESIGN AND SETTING

In this methodological review, we searched PubMed and Google Scholar from July 2009 to May 2022 for studies that used the hdPS for evaluating the effectiveness or safety of healthcare interventions. Two reviewers independently extracted study characteristics and assessed how the hdPS was applied and reported. Risk of bias was evaluated with the Rrisk Of Bias In Non-randomised Studies - of Interventions (ROBINS-I) tool.

RESULTS

In total, 136 studies met the inclusion criteria; the median publication year was 2018 (Q1-Q3 2016-2020). The studies included 192 datasets, mostly North American databases (n = 132, 69%). The hdPS was used in primary analysis in 120 studies (88%). Dimensions were defined in 101 studies (74%), with a median of 5 (Q1-Q3 4-6) dimensions included. A median of 500 (Q1-Q3 200-500) empirically identified covariates were selected. Regarding hdPS reporting, only 11 studies (8%) reported all recommended items. Most studies (n = 81, 60%) had a moderate overall risk of bias.

CONCLUSION

There is room for improvement in the reporting of hdPS studies, especially regarding the transparency of methodological choices that underpin the construction of the hdPS.

A cluster of autochthonous dengue transmission in the Paris region - detection, epidemiology and control measures, France, October 2023

A cluster of three confirmed autochthonous dengue cases was detected in October 2023 in the Val-de-Marne department neighbouring Paris, France. This marks the northernmost transmission of dengue in Europe reported to date. The epidemiological and microbiological investigations and the vector control measures are described. This event confirms the need for early case detection and response to contain dengue in Europe, especially given the 2024 Summer Olympic and Paralympic Games, when millions of visitors will visit the Greater Paris area.

A global multi-hazard risk analysis of road and railway infrastructure assets

Transport infrastructure is exposed to natural hazards all around the world. Here we present the first global estimates of multi-hazard exposure and risk to road and rail infrastructure. Results reveal that ~27% of all global road and railway assets are exposed to at least one hazard and ~7.5% of all assets are exposed to a 1/100 year flood event. Global Expected Annual Damages (EAD) due to direct damage to road and railway assets range from 3.1 to 22 billion US dollars, of which ~73% is caused by surface and river flooding. Global EAD are small relative to global GDP (~0.02%). However, in some countries EAD reach 0.5 to 1% of GDP annually, which is the same order of magnitude as national transport infrastructure budgets. A cost-benefit analysis suggests that increasing flood protection would have positive returns on ~60% of roads exposed to a 1/100 year flood event.

Spatial distribution has been rarely studied in global disaster risk models. Here the authors address damaged networked infrastructure at the asset level for a wider range of hazards and reveal a global Expected Annual Damages ranging from $3.1 to 22 billion with a particular vulnerability of transport infrastructure in Small Island Developing States.

MSNA during prolonged post-faint hypotension

BACKGROUND

Following tilt-induced syncope, blood pressure usually recovers rapidly after tilt back to the horizontal position. However, in some patients, hemodynamic recovery is delayed, a condition recently termed "prolonged post-faint hypotension" (PPFH). The mechanism is thought to be mediated by increased vagal outflow rather than exaggerated peripheral vasodilatation and sympathetic withdrawal. To date, no muscle sympathetic nerve activity (MSNA) recordings have been reported in this condition, so we aimed to confirm that neither vasodilatation nor MSNA withdrawal was responsible.

OBJECTIVES

To retrospectively select patients with satisfactory recordings of continuous BP and MSNA during tilt-induced syncope. To compare hemodynamic and MSNA profiles in patients with PPFH to patients with normal recovery (NR) after tilt-back.

METHODS

All patients were studied in Christchurch, New Zealand, between 1998 and 2008 using continuous arterial BP monitoring, and microneurographic recordings of MSNA from the right leg. Only patients with satisfactory BP and MSNA data throughout baseline, head-up tilt and presyncope were selected. Stroke volume (SV), cardiac output (CO), and total peripheral resistance (TPR) were derived using Modelflow. After baseline measurements, patients were tilted to the head-up 60° position and given GTN spray if asymptomatic after 20 min. Following the onset of presyncope, patients were tilted slowly back to the horizontal. PPFH was defined as systolic BP <85 mmHg for at least 2 min after tilt-back. Measurements were averaged at baseline, early tilt, presyncope, early and late recovery. Within-group comparisons were made between baseline and all other time points. Between-group comparisons were made over all time points.

RESULTS

Patients with PPFH (7 males, age 46 ± 5 years, n = 8) and with NR (8 males, age 47 ± 6 years, n = 8) were selected. Presyncope was provoked by GTN in 4/8 patients in each group. In both groups, MAP remained below baseline during early and late recovery: PPFH 84 ± 5 versus 51 ± 5 and 64 ± 5 mmHg (p = 0.001, p = 0.001); NR 104 ± 5 versus 83 ± 5 and 93 ± 5 mmHg (p = 0.001, p = 0.03). However, MAP and HR were lower in the PPFH group (p = 0.004, p = 0.023). During early recovery, CO remained below baseline only in the PPFH group (p = 0.001), whereas TPR remained constant in both groups. In both groups, all MSNA indices tended to remain above baseline levels during early and late recovery. PPFH 25 ± 2 increased to 31 ± 6 and 29 ± 4 bursts/min (p = 0.09, 0.02); NR 23 ± 3 increased to 33 ± 3 and 34 ± 3 bursts/min (p = 0.06, 0.01).

CONCLUSIONS

PPFH does not appear to be mediated by exaggerated vasodilatation or sympathetic withdrawal. Delayed recovery of cardiac output by increased vagal outflow is a more likely mechanism.

Capacity for resolution of Ras-MAPK-initiated early pathogenic myocardial hypertrophy modeled in mice

Activation of Ras signaling in cardiomyocytes has been linked to pathogenic myocardial hypertrophy progression and subsequent heart failure. Whether cardiomyopathy can regress once initiated needs to be established more fully. A 'tet-off' system was used to regulate expression of H-Ras-G12V in myocardium to examine whether Ras-induced pathogenic myocardial hypertrophy could resolve after removal of Ras signaling in vivo. Ras activation at weaning for 2 wk caused hypertrophy, whereas activation for 4 to 8 wk led to cardiomyopathy and heart failure. Discontinuing H-Ras-G12V transgene expression after cardiomyopathy onset led to improved survival and cardiomyopathy lesion scores, with reduced heart:body weight ratios, demonstrating the reversibility of early pathogenic hypertrophy. Activation of Ras and downstream ERK 1/2 was associated with elevated expression of proliferating cell nuclear antigen and cyclins B1 and D1, indicating cell-cycle activation and reentry. Coordinate elevation of broad-spectrum cyclin-dependent kinase inhibitors (p21, p27, and p57) and Tyr15 phosphorylation of cdc2 signified the activation of cell-cycle checkpoints; absence of cell-cycle completion and cardiomyocyte replication were documented by using immunohistochemistry for mitosis and cytokinesis markers. After resolution of cardiomyopathy, cell-cycle activators and inhibitors examined returned to basal levels, a change that we interpreted as exit from the cell cycle. Cardiac cell-cycle regulation plays a role in recovery from pathogenic hypertrophy. The model we present provides a means to further explore the underlying mechanisms governing cell-cycle capacity in cardiomyocytes, as well as progression and regression of pathogenic cardiomyocyte hypertrophy.

Inhibition of CREB function in mouse epidermis reduces papilloma formation

We used a double transgenic tetracycline system to conditionally express A-CREB, a dominant negative protein that prevents the DNA binding and function of cAMP-responsive element binding protein (CREB) family members, in mouse basal epidermis using the keratin 5 promoter. There was no phenotype in the adult. However, following a 7,12-dimethylbenz(a)anthracene (DMBA)/phorbol-12-myristate-13-acetate two-stage skin carcinogenesis experiment, A-CREB-expressing epidermis develop 5-fold fewer papillomas than wild-type controls. However, A-CREB expression one month after DMBA treatment does not prevent papilloma formation, suggesting that CREB functions at an early stage of papilloma formation. Oncogenic H-Ras genes with A-->T mutations in codon 61 were found in wild-type skin but not in A-CREB-expressing skin 2 days after DMBA treatment, suggesting that A-CREB either prevents DMBA mutagenesis or kills oncogenic H-Ras cells. In primary keratinocyte cultures, A-CREB expression induced apoptosis of v-Ras(Ha)-infected cells and suppressed the expression of cell cycle proteins cyclin B1 and cyclin D1. These results suggest that inhibiting CREB function is a valuable cancer prevention strategy.

Accelerated tumor formation in a fatless mouse with type 2 diabetes and inflammation

Epidemiologic studies show a positive association between obesity and cancer risk. In addition to increased body adiposity and secretion of fat-derived hormones, obesity is also linked to insulin resistance, type 2 diabetes, and chronic inflammation. We used the fatless A-ZIP/F-1 transgenic mouse to dissociate the relative role of each of these underlying factors in the development of cancer. These mice are unique in that they do not have white fat but do develop type 2 diabetes. In two cancer models, the classic two-stage skin carcinogenesis protocol and the C3(1)/T-Ag transgenic mouse mammary tumor model, A-ZIP/F-1 mice displayed higher tumor incidence, tumor multiplicity, and decreased tumor latency than wild-type mice. We examined circulating levels of adipokines, growth factors, and cytokines. As expected, adipokines (i.e., leptin, adiponectin, and resistin) were undetectable or found at very low levels in the blood of fatless mice. However, insulin, insulin-like growth factor-I, growth hormone, vascular endothelial growth factor, and proinflammatory Th2 cytokines, such as interleukin (IL)-1beta, IL-4, and IL-6, were elevated in A-ZIP/F-1 mice. Additionally, we examined multiple phosphorylated proteins (i.e., protein kinase B/Akt and ErbB2/HER-2 kinase) associated with cancer development. Results show that many of these phosphorylated proteins were activated specifically in the A-ZIP/F-1 skin but not in the wild-type skin. These findings suggest that adipokines are not required for the promotion of tumor development and thus contradict the epidemiologic data linking obesity to carcinogenesis. We postulate that insulin resistance and inflammation are responsible for the positive correlation with cancer observed in A-ZIP/F-1 mice.